Messages in this thread | | | Date | Mon, 18 May 2009 18:45:52 +0800 | From | Balbir Singh <> | Subject | Re: [RFC] Low overhead patches for the memory cgroup controller (v2) |
| |
* KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> [2009-05-18 19:11:07]:
> On Fri, 15 May 2009 23:46:39 +0530 > Balbir Singh <balbir@linux.vnet.ibm.com> wrote: > > > * KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> [2009-05-16 02:45:03]: > > > > > Balbir Singh wrote: > > > > Feature: Remove the overhead associated with the root cgroup > > > > > > > > From: Balbir Singh <balbir@linux.vnet.ibm.com> > > > > > > > > This patch changes the memory cgroup and removes the overhead associated > > > > with LRU maintenance of all pages in the root cgroup. As a side-effect, we > > > > can > > > > no longer set a memory hard limit in the root cgroup. > > > > > > > > A new flag is used to track page_cgroup associated with the root cgroup > > > > pages. A new flag to track whether the page has been accounted or not > > > > has been added as well. > > > > > > > > Review comments higly appreciated > > > > > > > > Tests > > > > > > > > 1. Tested with allocate, touch and limit test case for a non-root cgroup > > > > 2. For the root cgroup tested performance impact with reaim > > > > > > > > > > > > +patch mmtom-08-may-2009 > > > > AIM9 1362.93 1338.17 > > > > Dbase 17457.75 16021.58 > > > > New Dbase 18070.18 16518.54 > > > > Shared 9681.85 8882.11 > > > > Compute 16197.79 15226.13 > > > > > > > Hmm, at first impression, I can't convice the numbers... > > > Just avoiding list_add/del makes programs _10%_ faster ? > > > Could you show changes in cpu cache-miss late if you can ? > > > (And why Aim9 goes bad ?) > > > > OK... I'll try but I am away on travel for 3 weeks :( you can try and run > > this as well > > > tested aim7 with some config. > > CPU: Xeon 3.1GHz/4Core x2 (8cpu) > Memory: 32G > HDD: Usual? Scsi disk (just 1 disk) > (try_to_free_pages() etc...will never be called.) > > Multiuser config. #of tasks 1100 (near to peak on my host) > > 10runs. > rc6mm1 score(Jobs/min) > 44009.1 44844.5 44691.1 43981.9 44992.6 > 44544.9 44179.1 44283.0 44442.9 45033.8 average=44500 > > +patch > 44656.8 44270.8 44706.7 44106.1 44467.6 > 44585.3 44167.0 44756.7 44853.9 44249.4 average=44482 > > Dbase config. #of tasks 25 > rc6mm1 score (jobs/min) > 11022.7 11018.9 11037.9 11003.8 11087.5 > 11145.2 11133.6 11068.3 11091.3 11106.6 average=11071 > > +patch > 10888.0 10973.7 10913.9 11000.0 10984.9 > 10996.2 10969.9 10921.3 10921.3 11053.1 average=10962 > > Hmm, 1% improvement ? > (I think this is reasonable score of the effect of this patch) >
Thanks for the test, I have a 4 CPU system and I create 80 users, larger config shows larger difference at my end. I think even 1% is quite reasonable as you mentioned. If the patch looks fine, should we ask for larger testing by Andrew?
> Anyway, I'm afraid of difference between mine and your kernel config. > plz enjoy your travel for now :)
Sorry, I did not send you my .config, why do you think .config makes a difference? I think loading AIM makes the difference and I also made one other change to the aim tests. I run with "sync" linked to /bin/true and use tmpfs for temporary partition and 20*numnber of cpus for number of users.
If required, I can still send out my .config to you.
-- Balbir
| |